Norms mapping parameter

The norms mapping parameter controls whether OpenSearch stores normalization factors for a field. During relevance scoring, OpenSearch uses these factors to account for field length: a query term that matches a short field value contributes more to the score than the same term in a long value. For example, if a review field has norms enabled, a search for tent scores the review Great tent higher than a 15-word review that also contains tent. If norms are disabled, both reviews receive the same score.

Norms take disk space. For each field that has norms enabled, OpenSearch stores up to 1 byte for every document that contains the field. The actual size can be smaller when many values have similar lengths. Disable norms on fields that you use for filtering, sorting, or aggregations but not for relevance scoring, or on fields for which the value length should not affect relevance.

By default, norms is enabled on text fields, for which relevance scoring is important, and disabled on keyword fields, which are typically used for filtering.

Disabling norms on a field

The following request creates an index named products containing the description field of type text with norms disabled:

PUT /products
{
  "mappings": {
    "properties": {
      "description": {
        "type": "text",
        "norms": false
      }
    }
  }
}

To add a field with norms disabled to an existing index, use the following request:

PUT /products/_mapping
{
  "properties": {
    "review": {
      "type": "text",
      "norms": false
    }
  }
}

Changing norms on an existing field

You cannot enable norms on a field that has norms disabled. OpenSearch rejects the mapping update.

Do not disable norms on a field that already contains indexed values. OpenSearch accepts the mapping update, but afterward it rejects every document that contains a value for the field. Closing and reopening the index or deleting the existing documents does not resolve the issue. To disable norms on a field that contains data, create an index containing the new mapping and reindex your documents into it.